DEVELOPMENT OF A MASH TL-3 TRANSITION BETWEEN GUARDRAIL AND PORTABLE CONCRETE BARRIERS

Transcription

1 Duplication for publication or sale is strictly prohibited without prior written permission of the Transportation Research Board Paper No DEVELOPMENT OF A MASH TL-3 TRANSITION BETWEEN GUARDRAIL AND PORTABLE CONCRETE BARRIERS by Robert W. Bielenberg, M.S.M.E. Midwest Roadside Safety Facility University of Nebraska-Lincoln 130 Whittier Building 2200 Vine Street Lincoln, Nebraska Phone: (402) Fax: (402) rbielenberg2@unl.edu (Corresponding Author) Ronald K. Faller, Ph.D., P.E. Midwest Roadside Safety Facility University of Nebraska-Lincoln 130 Whittier Building 2200 Vine Street Lincoln, Nebraska Phone: (402) Fax: (402) rfaller1@unl.edu David Gutierrez, M.S.C.E. Survey Specialist VII Gyrodata, Inc Jefferson St. Houston, TX Phone : (308) david.gutierrez@gyrodata.com John D. Reid, Ph.D. Mechanical & Materials Engineering Midwest Roadside Safety Facility University of Nebraska-Lincoln W342 NH (0526) Lincoln, Nebraska Phone: (402) Fax: (402) jreid@unl.edu Phil Tenhulzen, P.E. Design Standards Engineer Roadway Design Division Nebraska Dept. of Roads Phone: (402) phil.tenhulzen@nebraska.gov Submitted to Transportation Research Board 96 th Annual Meeting January 8-12, 2017 Washington, D.C. Submission Date: November 15, 2016 Length of Paper: 5,994 (text) + 1,500 (6 figures) = 7,494words

2 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 2 ABSTRACT Often, road construction requires that work zones be created and shielded by portable concrete barriers (PCBs) to protect workers and equipment from errant vehicles as well as prevent motorists from striking other roadside hazards. For an existing W-beam guardrail system installed adjacent to the roadway and near the work zone, guardrail sections are removed in order to place a PCB system. The focus of this research study was to develop a crashworthy transition between W-beam guardrail and PCB systems. This research effort was accomplished through development and refinement of design concepts using computer simulation with LS-DYNA. Additionally, a Critical Impact Point (CIP) study was conducted to determine impact locations for full-scale crash testing. The design effort resulted in a new system consisting of a Midwest Guardrail System (MGS) which overlapped a series of F-shape PCB segments placed at a 15:1 flare. In the overlapped region of the barrier systems, uniquely-designed blockout supports and a specialized W-beam end shoe mounting bracket were used to connect the systems. Three full-scale vehicle crash tests were successfully conducted according to the AASHTO Manual for Assessing Safety Hardware (MASH) Test Level 3 (TL-3) safety performance criteria. Based on the successful test results, a MASH TL-3 crashworthy guardrail to PCB transition system is now available for protecting motorists, workers, and equipment found in work zones. Keywords: Crash Test, MASH, Portable Concrete Barrier, TL-3, Guardrail, Transition, MGS, F- shape

3 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 3 INTRODUCTION A transition between portable concrete barriers (PCBs) and W-beam guardrail is necessary when roadway construction requires that a work zone be created to shield workers and equipment from errant vehicles. For an existing W-beam guardrail system installed adjacent to the roadway and near the work zone, guardrail sections are removed in order to place the PCBs. The region where the two barriers meet can create a potential safety hazard if a proper transition is not installed, including the potential for poor vehicle capture, snag on the end of the PCBs, and rapid vehicle deceleration. The primary concerns associated with a connection between W-beam guardrail and PCBs correspond to the difference in barrier deflections and functionality of the barriers. Strong-post, W-beam guardrail systems have dynamic barrier deflections of approximately 40 in. for Test Level 3 (TL-3) impacts with passenger vehicles. However, free-standing PCB systems may have a dynamic barrier deflection as high as 80 in. under similar impact scenarios. While a transition from guardrail to PCBs may not need to be as stiff as a conventional approach guardrail transitions to bridge rails, it must provide sufficient lateral stiffness and strength to prevent pocketing as well as shield the end of the concrete barrier. Therefore, a proper transition in lateral barrier stiffness and strength is necessary to properly connect the two systems. Unfortunately, a crashworthy transition between guardrail and PCBs was unavailable. To address this need, the Nebraska Department of Roads (NDOR) sponsored research with the objectives of developing, testing, and evaluating a transition system between W-beam guardrail and free-standing PCBs that would improve safety for the motoring public and workzone workers. This transition was developed through a combination of design, computer simulation, and full-scale crash testing to meet the Test Level 3 (TL-3) safety performance criteria set forth in the AASHTO Manual for Assessing Safety Hardware (MASH) (1). The research focused on the F-shape PCB system developed through the Midwest States Pooled Fund Program and used by NDOR (2-3). BACKGROUND A literature review was conducted to identify research related to W-beam guardrails, PCB systems, and transitions in an effort to gain knowledge of barrier deflections and transitioning techniques. Individual barrier and transition system performance was reviewed in terms of vehicle snag, vehicle capture and stability, barrier pocketing, and barrier connection design. Full details regarding the literature search can be found in Gutierrez and Lingenfelter (4-5). From a review of the two barrier types, the F-shape PCB had almost two times the dynamic deflection of most W-beam guardrail systems. MASH TL-3 testing of the F-shape PCB had a maximum lateral dynamic barrier deflection of 79.7 in. (3). Similar testing of W-beam guardrail systems (6) found an average dynamic deflection of 39.7 in. and 41.4 in. for 27¾-in. and 31-in. tall guardrail systems, respectively. System behavior also differed for the two barriers. For W-beam guardrail, barrier deflections began immediately upon impact, and peak dynamic deflection typically occurred when the vehicle was parallel with the barrier. However, PCB deflections were slower to develop due to the inertia of the PCB segments, and the peak deflection of the F-shape PCB occurred after the vehicle exited the barrier system as the segments continued to slide on the pavement.

4 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 4 Design Criteria Basic design criteria were developed to meet NDOR needs regarding barrier performance and future implementation and included: 1. The transition should meet MASH TL-3 while minimizing barrier pocketing, vehicle snag, vehicle instability, and rapid deceleration. 2. The preferred transition would accommodate existing G4(1S) guardrail systems (7). However, the 31 in. tall Midwest Guardrail System (MGS) (8) was allowable if required. 3. Guardrail through reduced post spacing or PCB stiffening through anchorage was not desired. 4. The transition should consider reverse-direction impacts resulting from two-way traffic. 5. When transition installations occur near unpaved surfaces, PCBs will be installed on a compacted, crushed rock pad to mitigate barrier gouging in soil, tipping, and excessive rotation. DEVELOPMENT AND SIMULATION OF TRANSITION CONCEPTS A variety of transition concepts were formulated to meet the design criteria and ranked by their feasibility, potential safety performance, and ease of installation. It was preferred for the transition to be easy to install and limit the number of additional components. Therefore, each concept was presented in its simplest form, and additional features were added, as needed, to improve performance. A review of potential transition concepts focused the research to two preferred concepts, as described below and shown schematically in Figure Flared PCB Modified G4(1S) - Three 15H:1V flared PCB segments extended behind a modified G4(1S) guardrail system with posts interfering with installed PCB segments removed. 2. Parallel PCB Modified G4(1S) - Two PCB segments placed parallel to and behind a modified G4(1S) guardrail system before PCB system was flared at 15H:1V to create work zone. a) Flared PCB Modified G4(1S) b) Parallel PCB Modified G4(1S) FIGURE 1 Preferred Transition Concepts. Simulation of Transition Concepts LS-DYNA was used to analyze and refine the transition concepts (9). According to TL-3 of MASH, transitions must be impacted at a nominal speed and angle of 62 mph and 25 degrees,

5 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 5 respectively. Therefore, each candidate design was subjected to simulated impacts according to these parameters and at several impact locations ranging from the connection point between the guardrail and the PCB system to four posts upstream from the PCB system. Criteria used in the analysis of the concepts included vehicle behavior, occupant risk, and rail pocketing. Vehicle behavior was examined to evaluate the potential for vehicle capture and redirection without vehicle instability. Vehicle snag on the upstream end of the PCB system or other transition components could affect vehicle stability and cause rapid deceleration. Occupant risk measures, including occupant impact velocities (OIVs) and occupant ridedown accelerations (ORAs), were evaluated to determine the degree of hazard to occupants in the impacting vehicle. Finally, rail pocketing angle is a measure of the effective angle of the deformed barrier in front of the vehicle as it is redirected. Angles above 23 degrees were a concern for the transition as excessive pocketing angles have been shown to be associated with degraded barrier performance, including rail rupture (10). LS-DYNA models of the free-standing F-shape temporary concrete barrier, modified G4(1S), and MGS had been developed in previous studies (11-12). The Chevrolet Silverado vehicle model was chosen for the simulation study due to the likelihood of increased barrier deflections, rail and anchor loads, rail pocketing, and wheel snag with this vehicle type. Further, vehicle instabilities have been exhibited during full-scale crash tests involving 2270P pickup trucks with F-shape PCB systems due to vehicle climb. The Silverado vehicle model was originally created by the National Crash Analysis Center (NCAC) and later modified by MwRSF personnel. Concept No. 1 - Flared PCB-Modified G4(1S) Simulation Analysis of the transition concepts began with simulation of concept no. 1, the Flared PCB Modified G4(1S) concept. Several variations of the concept were analyzed, starting with a basic overlapping and connection of the flared PCB and the guardrail, as shown in Figure 1. Subsequent modifications were made to improve transition performance through the use of thrie beam, blockouts between the guardrail and PCB, addition of a cantilever beam off the end of the PCBs, and guardrail nesting. Full discussion of the simulations are found in Gutierrez (4), but several notable conclusions were drawn from the simulation of the flared PCB-modified G4(1S) concept. First, the modified G4(1S) system lacked the height and stiffness to safely capture and redirect the vehicle as problems with vehicle stability and barrier pocketing were noted at several impact points. A transition involving thrie beam upstream from the PCBs was simulated, which yielded improved vehicle stability. However, the guardrail support posts had a tendency to wedge against the PCBs, which led to elevated occupant risk values and rail pocketing angles. As such, posts were removed and replaced with blockouts attached to the PCBs. Blockouts were installed at a standard 6-ft 3-in. post spacing in the later configurations. The fully-blocked, thrie beam configuration yielded results with improved vehicle stability and occupant risk but with high rail pocketing angles. The pocketing behavior was caused by delayed displacement of the PCBs at the beginning of the impact event. Nested thrie beam was implemented to stiffen the guardrail system ahead of the PCB system and improve rail pocketing. The nested, fully-blocked, thrie beam rail configuration yielded improved pocketing angles. A comparison of simulation results for the baseline Flared PCB Modified G4(1S) concept and the improved design with nested, fully-blocked, thrie beam rail are shown in Figure 2.

6 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 6 a) Initial Flared PCB Modified G4(1S) Concept b) Flared PCB Modified G4(1S) Concept with Nested, Fully-Blocked Thrie Beam Rail sec sec sec sec sec Initial Concept Nested, Fully-Blocked Thrie Beam Rail c) Sequential Comparison FIGURE 2 Flared PCB Modified G4(1S) Simulation Configuration Comparison.

7 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 7 Concept No. 2 - Parallel PCB-Modified G4(1S) Simulation Analysis of the transition concepts continued with simulation of concept no. 2, the Parallel PCB Modified G4(1S) concept, which utilized modified G4(1S) guardrail attached to F-shape PCBs with two segments installed parallel to and behind the guardrail system. Modifications were implemented into the concept based on the simulation results from the Flared PCB Modified G4(1S) concept. Thus, nested thrie beam was installed for the final five rail sections in the transition adjacent to the PCBs, and the posts in front of the PCBs were removed and replaced with blockouts attached to the PCB segments at standard 6-ft 3-in. spacing. The nested thrie beam-with-fully-blocked rail configuration yielded two marginal longitudinal ORA values but had acceptable vehicle stability and rail pocketing angles. The simulation results for the Flared PCB-Modified G4(1S) and Parallel PCB-Modified G4(1S) concepts were reviewed, and the nested thrie beam configurations with fully blocked out rail for both the transition concepts were deemed to have potential to meet MASH TL-3. The parallel PCB configuration did not pose any discernable benefit as compared to a flared PCB configuration, so the flared configuration was preferred due to its reduced barrier overlap and simplicity. Concerns were raised that the incorporation of thrie beam elements may be overly complex and labor intensive. Therefore, it was recommended that the modified G4(1S) be replaced with the Midwest Guardrail System (MGS) as a third design concept. It was anticipated that the 31-in. top mounting height of the MGS would aid in vehicle capture and redirection without a transition to thrie beam. Concept No. 3 - Flared PCB-MGS Simulation The Flared PCB MGS concept was similar to the Flared PCB Modified G4(1S) concept, except MGS was connected to the 15H:1V flared PCB system in lieu of modified G4(1S). Although simulation results for the modified G4(1S) indicated that posts in front of PCBs would deform and wedge against the face of PCBs, the increased rail height of the MGS was believed to improve capture and redirection of the 2270P vehicle with reduced instabilities. Additionally, two posts remained in front of the first PCB segment and were intended to aid in PCB displacement. Upon impact, the posts were expected to rotate backward into the PCB and initiate displacement, which would reduce vehicle climb and instabilities. Simulation results for the baseline Flared PCB MGS configuration yielded high occupant risk values due to vehicle snag, and pocketing angles were a concern for impacts upstream from the PCB system. Thus, modifications to the configuration were investigated to improve performance. Modified configurations were simulated using blockout variations and cantilever beams, but high pocketing angles continued to be a concern with the attachment of the MGS to PCBs. In order to improve this behavior, a Flared PCB MGS configuration with nested rail placed upstream from and in front of the PCB system was simulated in order to stiffen the barrier system and lower pocketing angles. The simulation results for the nested-mgs configuration showed that occupant risk measures and pocketing angles were reduced to acceptable levels for all impact locations. Concept Selection The simulation results from the Flared PCB Modified G4(1S), Parallel PCB Modified G4(1S), and Flared PCB MGS concepts were compared and used to select the transition configurations

8 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 8 with the best performance. The compared metrics included, occupant risk values, vehicle orientation angles to determine relative stability, vehicle snag, and barrier pocketing angles. The minimum value for each metric represented the safest transition design. The metrics that exceeded or were within 20 percent of MASH or recommend limits, practicality, and ease of installation were used to rank the configurations within each design concept as well as to establish whether each configuration had a high, moderate, or low likelihood of success. The comparison and rankings are shown in detail in Gutierrez (4). Based on the rankings, the Flared PCB MGS with nested rail configuration was preferred. It was the only configuration within all three concepts that did not display concerns for degraded vehicle behavior and occupant risk, nor show high pocketing angles. Also, nesting of MGS would not significantly increase installation effort as compared to several other promising configurations, including a transition to thrie beam. Thus, the Flared PCB MGS configuration with nested rail was selected as the preferred alternative and recommended for full-scale crash testing and evaluation. SELECTION OF CRITICAL IMPACT POINTS AND TEST MATRIX According to TL-3 of MASH, transitions between longitudinal barrier systems must be subjected to two full-scale vehicle crash tests: 1. Test No Impact with 1100C vehicle at Critical Impact Points (CIP) of transition system at 62 mph and 25 degrees. 2. Test No Impact with 2270P vehicle at CIP of the transition system at 62 mph and 25 degrees. It was also recommended that an additional reverse-direction impact of test designation no with the 2270P vehicle be required for evaluating the transition as it may be subjected to twoway traffic adjacent to the barrier. Because the transition design did not include stiffening of the two semi-rigid barrier systems as they approach one another, it was believed that separate evaluation of the stiffness transition and the barrier connection point was not warranted. Computer simulation was conducted in order to determine the CIPs. This analysis consisted of using LS-DYNA simulation to select the critical attachment point between the MGS and PCB systems and the CIP for test no with the 2270P vehicle for both oncoming and reverse-direction traffic. Full discussion of the CIP analyses is detailed in Gutierrez (4), but the relevant conclusions were: 1. The 2270P CIP for the transition from guardrail to PCB was identified as the center of the fifth guardrail post upstream from the W-beam end shoe attachment based on this location generating the highest barrier pocketing. 2. The 2270P CIP for a reverse-direction impact into the transition from PCB to guardrail was identified as 12.5 ft upstream from the W-beam end shoe connection to the PCB. This point was selected as it maximized the climb of the 2270P vehicle on the face of the PCB segment and caused potential concerns with vehicle capture on the W-beam rail as it traversed the system. Engineering analysis and review of previous MASH testing with the 1100C vehicle were used to select a CIP for test no Maximizing vehicle extension under the guardrail and simultaneous interactions with the PCB in order to promote wedging of the corner of the small car under the guardrail and between the two overlapping barrier systems were considered. This type of behavior promotes potential increases in vehicle deceleration, instability, and loading on the guardrail element. Previous testing of an MGS approach guardrail transition with a 4-in. tall

9 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 9 wedge-shaped curb has demonstrated that combined loading caused by the front corner of the vehicle being wedged vertically between the curb and the guardrail was sufficient to result in rail rupture (13). Review of this approach guardrail transition and other full-scale crash tests indicated that the CIP selected for test no was located 93¾ in. upstream from the second guardrail splice upstream from the end shoe connection. This ensured that the vehicle critically loaded a splice while being engaged with W-beam guardrail and PCBs. Additionally, this CIP evaluated a potential for non-desirable vehicle interaction with the W-beam end shoe mounting bracket. FINAL TRANSITION DETAILS The final transition was comprised of a tangent, nested-mgs that overlapped an adjacent, PCB system oriented at a 15:1 flare, as shown in Figure 3. Minimum installation recommendations for testing and evaluating the transition were based on the initial computer simulation analysis and consisted of the following: 1. For testing purposes, the transition should consist of at least a ft long MGS system and an eleven segment PCB system positioned at a 15H:1V flare. 2. The transition required a minimum of three PCB segments extending behind the nested MGS at a 15:1 flare, which corresponded to guardrail attachment to the upstream end of the fourth PCB segment. Additional PCBs flared behind the MGS would not be an issue as the potential for vehicle and barrier interaction with the PCBs was maximized for the minimum overlap condition. 3. Placement of standard MGS posts and blockouts was not recommended within the first two sections of guardrail upstream from the W-beam end shoe connection as the PCBs would interfere with existing posts. Thus, connection between the guardrail and the PCB segments in that region was to be accomplished with specially-designed blockout mounts. 4. A minimum of five 12-ft 6-in. long, nested W-beam sections were utilized upstream from the end-shoe connection to the PCB. For the minimum PCB overlap noted above, this corresponds to one complete 12.5-ft long section of nested rail upstream from the end of the PCBs. The MGS was constructed with sixteen steel posts spaced at 75 in. on center. The line posts were W6x8.5 sections with an embedment depth of 40 in. A 6-in. wide x 12-in. deep x 14¼-in. long blockout was used to block the rail away from the front face of each steel post. The 12-gauge W-beam was mounted at a height of 31 in. and nested for the first five 12-ft 6-in. long rail sections upstream of the W-beam end shoe. A tangent anchorage system was utilized on the upstream end of the MGS. Eleven 12-ft 6-in. long F-shape PCBs were connected to the MGS system using a stiffness transition. The concrete barriers were 22½ in. wide at the base and 8 in. wide at the top. Each barrier segment was interconnected by 1¼-in. diameter, A36 steel connection pins and connector plates placed between ¾-in. diameter reinforcing loop bars extending from the end of the barrier sections. All PCB segments were set on a 6-in. deep compacted crushed limestone pad meeting AASHTO Grade B soil specifications or on the concrete tarmac. The overlapped portion of the transition from MGS to PCB incorporated four blockouts between the guardrail and concrete barrier nos. 2 and 3 mounted on bent plate blockout attachments. The bent plate blockout attachment accounted for the vertical flare of the PCB, and individual timber blockouts were then cut on one face to match the offset depth and 15:1 flare of the PCB segments. Although the mounting plate had four holes, it was secured to the PCB using

10 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 10 only two ¾-in. diameter x 6-in. long Powers Wedge Bolts. The additional holes provide an alternative for improper anchor installation or rebar interference. The bracket allowed for guardrail to be bolted to the blockout using guardrail bolts. The guardrail was connected and transitioned to the concrete barrier at an angle of 3.8 degrees by a steel mounting bracket and W-beam end shoe. The basic design of the W-beam end shoe mounting bracket was similar to attachments that have been previously developed for attachment of thrie beam approach guardrail transitions to sloped concrete parapets. The steel mounting bracket was mounted on the impact side of the fourth PCB segment and 10 in. away from the upstream end with four 1-in. diameter A325 Grade A bolts. The downstream end of the bracket was angled 8.0 degrees to be flush against the concrete barrier. A W-beam end shoe was attached to the front side of the connector plate with five ⅞-in. diameter A325 bolts secured by A563 nuts welded to the interior of the connection plate.

11 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen FIGURE 3 Guardrail to PCB Transition System. 11

12 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 12 TEST NO. MGSPCB-1 (MASH 3-21) In test no. MGSPCB-1, a 4,914-lb pickup truck impacted the MGS to PCB transition at a speed of 63.2 mph and at an angle of 25.3 degrees, as shown in Figure 4. Initial vehicle impact occurred 2.5 in. downstream from the fifth guardrail post upstream from the W-beam end shoe. The vehicle was captured by the W-beam rail element and redirected. No vehicle snag on the PCB system was observed. The vehicle snagged on the second and third posts upstream from the W-beam end shoe due to post deflection backward and against the first PCB segment, but the vehicle continued to be safely redirected. At sec after impact, the vehicle became parallel to the barrier. At sec, the vehicle exited the system. The vehicle came to rest 234 ft 1 in. downstream from impact and 21 ft 11 in. in front of the barrier, and its trajectory did not violate the bounds of the exit box. Barrier damage was moderate and consisted of rail deformation, damaged timber blockouts, bending of steel posts, contact marks on the front face of the concrete segments, and spalling of the concrete, as shown in Figure 4. Five of the guardrail posts were deformed, and the third post downstream of the impact point was twisted and bent downstream with the downstream side of the post against the upstream face of first PCB segment. Contact marks from the vehicle were visible on the front face of the first two PCB segments, but no contact was noted with the upstream end of the first PCB. The blockout mounts and the W-beam end shoe mounting bracket were undamaged. The maximum lateral dynamic deflections were 36.1 in. for the rail, 27.7 in. at the first post downstream of impact, and 6.7 in. at the downstream end of first PCB segment. The working width of the system was 58.7 in. Exterior vehicle damage was moderate and concentrated on the right-front corner and right side of the vehicle where the impact occurred, as shown in Figure 4. The interior occupant compartment deformations were minimal and did not violate the limits established in MASH. Longitudinal and lateral occupant impact velocity (OIV) were 12.8 ft/s and 15.7 ft/s, respectively. The longitudinal and lateral occupant ridedown acceleration (ORA) were g s and g s, respectively. The longitudinal ORA of g s occurred due to vehicle snag on system posts that were deflected against the first PCB segment. Vehicle stability was acceptable. Test no. MGSPCB-1 was determined to be acceptable according to the TL-3 safety performance criteria found in MASH.

13 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen sec sec sec sec (b) Barrier Damage sec sec (a) Sequential Events (e) Vehicle Damage FIGURE 4 Sequential Photographs and Barrier and Vehicle Damage, Test No. MGSPCB-1.

14 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 14 TEST NO. MGSPCB-2 (MASH 3-20) In test no. MGSPCB-2, a 2,436-lb car impacted the PCB to MGS transition at a speed of 65.1 mph and at an angle of 24.0 degrees, as shown in Figure 5. Initial vehicle impact occurred 99.5 in. upstream from the centerline of the second splice upstream from the W-beam end shoe. The vehicle was initially captured by the W-beam rail element and began to be redirected. The vehicle bumper, right-front fender, and right-front tire extended under the W-beam rail and impacted the second PCB segment in the system at sec after impact, but the vehicle continued to be safely redirected as it engaged the two overlapped barrier systems. At sec after impact, the vehicle became parallel to the barrier. At sec, the vehicle exited the system. The vehicle came to rest 157 ft 5 in. downstream from impact and 22 ft in front of the barrier oriented downstream, and its trajectory did not violate the bounds of the exit box. Barrier damage was moderate and consisted of rail deformation, damaged timber blockouts, contact marks on the front face of the concrete segments, and spalling of the concrete barriers, as shown in Figure 5. The blockout mounts and the W-beam end shoe mounting bracket were undamaged. The maximum lateral dynamic deflections were 26.3 in. for the rail, 3.1 in. at the first post upstream from impact, and 28.1 in. at the downstream end of the second concrete barrier segment. The working width of the system was 61.4 in. Exterior vehicle damage was moderate and was concentrated on the right-front corner and right side of the vehicle where the impact occurred, as shown in Figure 5. The windshield was deformed and shattered and had a 23-in. long tear at the top located 10 in. from the left A-Pillar, caused by deployment and contact from the front, passenger airbag with the windshield. Because the damage was not due to vehicle interaction or direct contact with the barrier system, it was not considered in the test evaluation. The interior occupant compartment deformations were minimal and did not violate the limits established in MASH. Longitudinal and lateral OIV values were ft/s and ft/s, respectively. The longitudinal and lateral ORA values were g s and g s, respectively. Vehicle stability was acceptable. Test no. MGSPCB-2 was determined to be acceptable according to the TL-3 safety performance criteria found in MASH.

15 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen sec sec sec sec (b) Barrier Damage sec sec (e) Vehicle Damage (a) Sequential Events FIGURE 5 Sequential Photographs and Barrier and Vehicle Damage, Test No. MGSPCB-2.

16 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 16 TEST NO. MGSPCB-3 (MASH 3-21, REVERSE DIRECTION) In test no. MGSPCB-3, a 5,012-lb pickup truck impacted the PCB to MGS transition at a speed of 63.1 mph and at an angle of 24.6 degrees, as shown in Figure 6. The vehicle impacted 12 ft 9 in. upstream from the centerline of the W-beam end shoe on the fifth PCB segment in the system and began to be redirected. By sec, the right-front fender contacted the leading edge of the W-beam end shoe mounting bracket, and the vehicle began to interact with the W- beam rail. A portion of the right-front fender and right door snagged on the mounting bracket, but the vehicle continued to be safely redirected. At sec after impact, the vehicle became parallel to the barrier. At sec, the vehicle exited the system. The vehicle came to rest 187 ft 9 in. downstream from impact and 56 ft 10 in. behind the barrier oriented with the front of the vehicle facing away from the back side of the barrier, and its trajectory did not violate the bounds of the exit box. Barrier damage was moderate and consisted of cracking of the concrete, contact marks on the front and top face of the concrete segments and the face of the W-beam rail, spalling of the concrete, damaged timber blockouts, and W-beam rail deformation, as shown in Figure 6. The first two impacted PCB segments displayed vertical cracking on the front and back faces of the barriers and minor concrete spalling. Only minor damage was noted to the remaining PCB segments. The blockout mounts and the end shoe mounting bracket were undamaged, except for minor scuff marks on the end shoe mounting bracket. The maximum lateral dynamic deflection of the rail and concrete barriers for the system was 30.6 in. for the rail and 37.2 in. at the upstream target on second impacted concrete barrier segment. Guardrail post motion was negligible. The working width of the system was 58.7 in. Exterior vehicle damage was moderate and concentrated on the right-front corner and right side of the vehicle where the impact occurred, as shown in Figure 6. Deformation and tearing of the right-front fender and right-side doors occurred due to snagging on leading edge of the W-beam end shoe mounting bracket. The right fender was bent upward 9 in. from the top edge of the wheel well, starting at the back of the fender and extending 20 in. forward. The rightfront door had a 23-in. tall by 15-in. wide tear located at the front, 11 in. from the bottom, while the right-rear door had an 8½-in. long by 3-in. tall tear located 17 in. from the bottom. The tears in the door were to the exterior sheet metal only and did not compromise the occupant compartment. The interior occupant compartment deformations were minimal and did not violate the limits established in MASH. Longitudinal and lateral OIV values were ft/s and ft/s, respectively. The longitudinal and lateral ORA values were g s and g s, respectively. Vehicle stability was acceptable. Test no. MGSPCB-3 was determined to be acceptable according to the TL-3 safety performance criteria found in MASH.

17 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen sec sec sec sec (b) Barrier Damage sec sec (e) Vehicle Damage (a) Sequential Events FIGURE 6 Sequential Photographs and Barrier and Vehicle Damage, Test No. MGSPCB-3.

18 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 18 SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS A crashworthy transition between guardrail and free-standing PCB was designed, tested, and evaluated under the safety requirements for MASH TL-3. The guardrail to PCB transition system was developed using extensive LS-DYNA simulation that investigated and refined potential concepts. Concepts were modified to enhance vehicle stability and capture as well as mitigate occupant risk, vehicle snag, and barrier pocketing. Concept refinement led to a transition system comprised of a tangent, nested-mgs that overlapped an adjacent, flared PCB system. LS-DYNA simulation was also used to identify critical impact points for full-scale crash testing. The transition system was subjected to three full-scale crash tests and successfully evaluated according to TL-3 impact safety standards provided in MASH. These tests evaluated structural integrity, vehicle snag, vehicle instability, and vehicle capture. The MASH TL-3 transition provides the first crashworthy option for the connection of MGS and F-shape PCBs. The transition design should be easy to implement as it requires minimal alterations of the guardrail and PCBs. As with any new barrier system, the guardrail to PCB transition needed to consider implementation guidance and provide recommendations for real-world installations. The recommended minimum system configuration for real-world installations is: 1. A minimum ft long MGS system and eleven, 12.5-ft long, F-shape PCB segments at a 15H:1V flare should be used. A minimum of eight PCBs should be placed downstream from the point where the W-beam guardrail attaches to PCBs. Shorter lengths for either barrier would need to be further evaluated. 2. The transition requires a minimum of three PCB segments extending behind the nested MGS at the 15H:1V flare, which allows anchorage of the guardrail to the upstream end of the fourth PCB segment. Additional length of overlapped, flared PCB is acceptable. 3. In order to provide adequate anchorage for the end shoe mounting bracket on the PCB, the anchor bracket mounting bolts that extend through the PCB must be mounted a minimum of 12 ¼ in. away from the upstream edge of the PCB segment similar to that used in the full-scale crash testing detailed herein. 4. A minimum of five 12-ft 6-in. long, nested W-beam sections must be utilized upstream from the end-shoe connection to PCBs. 5. The 15H:1V flare used in the transition to offset PCBs behind the guardrail will likely convert to PCBs tangent to the roadway once the work zone area has been established. It was recommended that conversion from the 15H:1V flare to tangent to the roadway not begin until a minimum of two PCB segments downstream from the W-beam end shoe connection. Additional recommendations regarding grading, surfacing, and clear areas behind the transition as well as system repair guidance are provided in Lingenfelter (5). Finally, the guardrail to PCB transition developed herein focused on the MGS guardrail system and 12.5-ft long, F-shape PCBs. While the transition was designed specifically for these two barrier systems, there may be a desire to integrate this transition with other barrier systems, including existing G4(1S) W-beam guardrail or alternative PCB designs. Guidance is provided in Lingenfelter (5) regarding transitioning from existing G4(1S) systems to the MGS in advance of the transition. Additionally, it is believed that the transition could be adapted to other PCB systems with considerations for barrier segment capacity, joint design, barrier geometry, and other factors. However, further research and testing would likely be required to evaluate safety performance.

19 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 19 ACKNOWLEDGMENTS The authors wish to acknowledge the Nebraska Department of Roads for sponsoring and guiding the project. The simulation effort was completed utilizing the Holland Computing Center of the University of Nebraska. REFERENCES Manual for Assessing Safety Hardware (MASH), American Association of State Highway and Transportation Officials (AASHTO), Washington, D.C., Faller, R.K., Rohde, J.R., Rosson, B.T., Smith, R.P., and Addink, K.H., Development of a TL 3 F Shape Temporary Concrete Median Barrier, Project SPR 3(017), Report No. TRP , Midwest Roadside Safety Facility, UNL, December Polivka, K.A., Faller, R.K., Sicking, D.L., Rohde, J.R., Bielenberg, B.W., Reid, J.D., and Coon, B.A., Performance Evaluation of the Free-Standing Temporary Barrier - Update to NCHRP 350 Test No with 28" C.G. Height (2214TB-2), NCHRP, Report No. TRP , Midwest Roadside Safety Facility, UNL, October 12, Gutierrez, D.A., Bielenberg, R.W., Faller, R.K., Reid, J.D., and Lechtenberg, K.A., Development of a Mash TL-3 Transition Between Guardrail and Portable Concrete Barriers, Report No. TRP , Midwest Roadside Safety Facility, UNL, June 26, Lingenfelter, J.L., Kohtz, J.E., Bielenberg, R.W., Faller, R.K., and Reid, J.D., Testing and Evaluation of MASH TL-3 Transition between Guardrail and Portable Concrete Barriers, Draft Report No. TRP , Midwest Roadside Safety Facility, UNL, August Abu-Odeh, A.Y., Kim, K.M., and Bligh, R.P, Guardrail Deflection Analysis, Phase I: ( ), Research Report No , Texas Transportation Institute, College Station, TX, August Bullard, Menges, and Alberson, NCHRP Report 350 Compliance Test 3-11 of the Modified G4(1S) Guardrail with Timber Blockouts, Report No. FHWA-RD , Texas Transportation Institute, September Faller, Polivka, Kuipers, Bielenberg, Reid, Rohde, and Sicking, Midwest Guardrail System for Standard and Special Applications, Transportation Research Record No. 1890, Transportation Research Board of the National Academies, Washington, D.C., Hallquist, J.O. LS-DYNA Keyword User s Manual, Livermore Software Technology Corporation, Livermore, California, Polivka, K.A., Faller, R.K., Reid, J.D., Sicking, D.L., Rohde, J.R., and Holloway, J.C., Crash Testing of Missouri s W-Beam to Thrie Beam Transition Element, Report No. TRP , Midwest Roadside Safety Facility, UNL, September Bielenberg, R.W., Quinn, T.E., Faller, R.K., Sicking, D.L., and Reid, J.D., Development of a Retrofit, Low-Deflection, Temporary Concrete Barrier System, Report No. TRP , Midwest Roadside Safety Facility, UNL, Lincoln, Nebraska, March 31, Julin, R.D., Reid, J.D., Faller, R.K., and Mongiardini, M., Determination of the Maximum MGS Mounting Height Phase II Detailed Analysis Using LS-DYNA, Report No. TRP , Midwest Roadside Safety Facility, UNL, December Winkelbauer, B.J., Putjenter, J.G., Rosenbaugh, S.K., Lechtenberg, K.A., Bielenberg, R.W., Faller, R.K., and Reid, J.D., Dynamic Evaluation of MGS Stiffness Transition with

20 Bielenberg, Guttierrez, Faller, Reid, and Tenhulzen 20 Curb, Report No. TRP , Midwest Roadside Safety Facility, UNL, Lincoln, Nebraska, June 30, 2014.